基于GA-BP人工神经网络的农宅供热负荷预测技术研究
Research on rural house heating load prediction technology based on GA-BP artificial neural network
摘要：
农宅实际供热负荷的预测对农宅供暖系统优化和新型供暖系统的应用具有重要意义。本文基于对华北地区农宅室内外13个参数的长期监测和相关性分析,研究了经遗传算法优化的BP人工神经网络模型(GA-BPANN)应用于农宅供热负荷预测的可行性和可靠性。结果表明,当GA-BPANN输入变量为按与供热负荷相关性强度排序的前6个参数(室内温度、室外温度、室内TVOC(总挥发性有机化合物)浓度、室内相对湿度、室外相对湿度和光照强度)时可以得到高精度的预测结果,为合理确定预测方案和农宅供热负荷提供了借鉴。
Abstract：
The prediction of the actual heating load of the rural houses has an important significance for the optimization of the rural house heating system and the application of new heating systems.Based on the long-term monitoring and correlation analysis of 13 indoor and outdoor parameters of rural houses in the North China,this paper studies the feasibility and reliability of the BP artificial neural network model optimized by the genetic algorithm (GA-BPANN) applied to the heating load prediction of rural houses.The results show that high-precision prediction results can be obtained when the GA-BPANN input variables are the top 6 parameters ranked according to the heating load related intensity(indoor temperature,outdoor temperature,indoor TVOC concentration,indoor and outdoor relative humidity and light intensity),which provides a reference for the reasonable determination of the prediction scheme and the rural house heating load.
Keywords:rural house heating; heating system; heating load prediction; correlation analysis; artificial neural network